A New Research Tool For Environmental Studies LUT map : Land Use and Topography

A New Research Tool For Environmental Studies LUT map : Land Use and Topography

L. Kithsiri Perera and Ryutaro Tateishi
Remote Sensing and Image Research Center,
Chiba University, 1-33, Yayoi-cho,
Inage-ku, Chiba, JAPAN

Abstract
Some of the recent developments in Digital Mapping and GIS technology were used to produced a new thematic map. The new map uses different colors and tones to represent Land Use and Topographic (LUT map) features Simultaneously. A test study was conducted using Landsat MSS and Topographic data of southern Sri Lanka. Resulted map and the LUT data vase explain the distribution of five and categories under eight topographic levels, pictorially and numerically. As an example; Open Water area covers only 0.5% of the total Land and 95% of that belongs to area below 91m and almost no water was identified over the 275m level. This kind of detail data structures and the LUT map can be practically more useful in Land suitability assessments and Environmental management programs.

1. Introduction
Maps have long been a major source of information for planners and environmental researchers. Generally, data for making maps are gathered by ground surveys and use aerial photographs to update. With the beginning of the space era, Satellite Remote Sensing has been recognized as an useful, cost effective source of data for maps. At the same time the recent developments in Digital Mapping and other technical devisers have been accelerated the usage of satellite data together with other conventional data sources for the studies which is based on the Geographical Information System (GIS) (Yeh 1991 p.5). scientists believe the digital computer has had a much effect on maps which is equal to influence that made by the intention of the printing press or photograph on map production (Monmonier 1982 p.2). Digital map production procedure may start with the “capturing” or by digitizing information from various maps and image. I digitized data set is converted to numerical form and stored as a data base. This database is easy to store, edit, update and convert to maps and diagrams and these factors are associated with the main advantages of the digital mapping (Mather 1991 p.100).

Present day advancements of the digital mapping and GIS are well performed in developed world but give less benefits tot eh third world due to economical and technological reasons. A new concept of Appropriate Technology (AT) has been introduced to regional development programs in developing countries although GIS tools are still an essential part of regional development through the use of AT (Yapa 1991, p.40).

Most of the developing countries are still using 1 inch to 1 mile conventional Topographic map in their planning requirements and these maps have a greater accuracy in location but less information on land use types. On the other hand many of these countries are available with satellite data. In this research paper, some of the above discussed technological developments were utilized together with few conventional data sources and methods in order to construct a new thematic map. The procedure of the research uses Landsat MSS data of Southern Sri Lanka with 1:63,360 topographic maps of the area. The resulted map will show land use and topographic features simultaneously by colors and the map can be called as “LUT Map” since it is based on Land Use and Topographic data.

2.Study area and the Data base
The area which was selected for the study lies between two major rivers in southern Sri Lanka and contains approximately 1460 km² land area (See fig. 1). Eastern part of the area is relatively dry and the mean annual temperature is ranges from 22.5 – 27.5 C. Rainfall has greater differences within the area and the mean annual rainfall from 1000mm to 4000mm. Wet western half has much human activities including homestead gardens, spice and other plantation activities. Thick forest areas are mainly located in north western mountains and these are few isolated forest patches. Since the eastern part is relatively dry, these are number of major irrigation projects around that area. This physical and human structure was the key reason for selecting this particular area for the study.

Fig.1. Location and general information of the study area.
CCT of Landsat MSS data dated 16th March 1987 (Path 141; Row 5.5) and 1:63,360 scaled two topographic map sheets (Morawaka – 1943, and Ambalantota – 1995) were used for the study. Apart from that few Aerial Photographs (1982) were utilized as ground truth data.

3. Methodology
The overall production procedure of the LUT map is explained in the Fig. 2. It contains both digital and manual data processing steps. The final LUT map has to be produced through a full color printer or optionally the user may print each land cover class in separate sheets in back.

3.1 Construction of the Digital Elevation Model (DEM)
Basically, using a grid system height values can be collected from the Topographic map but a different method was followed to construct the DEM due to few reasons, i.e. 1) In the final stage these values have to be reclassified into less number of classes, 2) Area is considerably large to use a grid overlay.

Fig. 2 Production Procedure of the LUT map
Since it may cost time and labor, 3) Equal pixel size in DEM and Land use data file will produce a smooth result. Therefore different elevation zones have been reclassified (by level slicing), digitized and continually use the name; the DEM data file. In generating DEM, the study area was delineated on the topographic sheets and selected a rectangular are which touches the edges of the study area. The Topographic maps were relatively old productions hence the contours had been drown with 100 feet interval. A set of contour lines was delineated using a light table and these selected contour lines are able to show a remarkable height difference and divides the area in to 8 topographic levels as : 0 – 100 feet, 101 – 300, 301 – 500, 501 – 700, 701 – 900, 901 – 1200, 1201 – 1500 and over 1500. The value of the each class was converted in to meters.

Then the Black line map was digitized by a canner in to the computer and re-touched each line, to contain only one pixel width. On the screen 8 different colors were applied to each height polygons while including black lines into each class. This produces DEM a data file that only represents topographic classes in colors without boundary lines (See Fig. 3). Finally the DEM file was reduced to the same size of MSS data file which contains 764 pixels and 353 lines.

Fig 3. Digitized and level sliced Topography map - DEM data file of the study area
3.2 Land use map
Image processing was carried out by IDRISI (V. 4)and other available software facilities in the center. CCT of the Landsat MSS was red by a SUN Work Station and a sub image (1200 pixels x 700 lines) was selected from the Band 1,2,3 and 4. RGB color composites and image enhancements were carefully observed and 26 Ground Control Points (GCP) were selected with references on the topographic maps. Some of GCPs were rejected in the test procedure and final geometric correction was performed with 21 GCPs (R.M.S error = 0.30 in pixel, 0.32 in lines) and all image bands were rectified for the Transverse Mercater Projection. Then the study area was selected (764 pixels x 353 lines)and observed in detail by producing Principal Component Images and various color composites. Training sites were selected with the use of few air photos, maps and other published materials as well as characteristics of band histograms. By the supervised classification methods of Maximum Likelihood, all pixels in four image bands were classified into five Land use categories; i.e. 1) Forest, 2 ) Scrub, 3) Mixed vegetation, 4) Paddy and 5) Water and the resulted map is presented by Fig. 4.

Fig .4 Land use map (based on Digitally classified Landsat MSS data).
3.3 Preparation of the LUT Map
In this process land use and DEM data files will overlay together and for that purpose both data files have to be equal in the size. Land use map which was produced from MSS data has its full pixel out put and the DEM file was reduced to meet the size of Land use map. LUT map production steps can be explained as follows.

LUT map will use different colors to represent the different land use categories and various tones of the colors to indicate the elevation of each pixel. Because of that a well managed color scheme must be worked out. A combination of 7 or 8 color with maximum of 10 tones will be able to identify practically when the LUT map is in use. In this study five colors and eight tones of each color were selected to represent the land use and elevation categories.
DEM data file was reclassified in to value 1, 2, 3, 4, 5, 6, 7, 8 and area beyond the east and west side rivers was classed as 0.
In the final Land Use map, each land use category was assigned to new value as; 1(Forest, 9 (Scrub), 10 (Mixed Veg.), 11 (Paddy), 13 (Water).
These re-classed DEM and Land Use files were multiplied to produce the LUT map (DEM x Land SUE data file = LUT data file). Since the out side the study area in DEM file has been assigned to 0, the out side are of the LUT map also automatically categorized as 0 and each pixel within the area will contain both, land use and topographic information.
In the final step pixel in LUT data file have assigned to the values of the color scheme file. The outlook of the map highly depends on the color selection in the color scheme.

4. Advantages and limitations of the LUT map
By the Fig. 5, the LUT map of Morawaka area, is presented and the legend shows the land Use and Elevation classes. The main benefit of the LUT map is its ability to show the land use and elevation of any location on one map with an acceptable accuracy. The entire map shows a 3 dimension view of the area. The LUT map presents the land use changes according to the elevation and through its’ data base it is possible to calculate the distribution of any land use type among various topographic levels. The table 1 presents the numeric results of the study and shows how the five Land Use classes are distributed among the eight Topographic levels. This kind of analysis which is based on timely different data will give a vast amount of information to the environmental planners.

Fig 5. LUT map of the study area.
The final LUT data file is a database and it can be edited, analyzed or cooperated with other data sources with les time and cost. Since the elevation is not as contours, other linear features can be reprinted on the map with a color that is not in the map legend. As is shown in Fig. 6 the some of the land use categories of the Morakawa area has been analyzed. The land use changes according to the elevation are figured here. It is easy to identify not only the pattern of the land use changes but also the amount of the land use change in each height class. The numeric results have not been analyzed in detail but the results indicate multipurpose usages of the LUT database. If the DEM data were gathered as point values it can be converted to the Slope data and its is possible to produce a Land Use and Slope map by the same procedure of LUT map production. These information will benefit the environmental planning programs. As an example, according to the amount of the forest that changes in different heights and slopes will highly effect the entire river basin. The pattern and the degree of such kind of changes can be easily worked out by multi temporal data based LUT maps.
Table 1 – Area under the each land use and elevation category (in hectares)

Topo. Cla.	Forest	Scrub	Mixed Veg.	Paddy	Water	Sub Total
0 – 30	1256	13475	17231	9499	342	41803
31 – 91	1689	23838	21179	5732	426	52864
92 – 152	1858	10076	13528	1709	27	27198
153 – 213	1608	2530	3721	554	1	8414
214 – 274	1327	2065	2214	328	1	5935
275 – 366	1894	814	2057	268	0	5033
367 – 457	849	350	826	72	1	2098
Over 457m	640	1794	232	13	0	2679
Sub total	11121	54942	60988	18175	798	146024

Fig 6. Area under three Land Use types according to different Topographic levels.
Apart from above advantages there are some limitations in the LUT map. Maximum number of colors and tones is the main controlling factor for the amount of the information that can be stored on the map. This is due to the limitations in human ability to identify different colors and tones on one sheet. On the other hand new pixel values of the LUT map must not exceed 255 and because of that number of land use and elevation classes are inter-dependable. Personal skill in elevation data gathering and image interpretations also may influence the accuracy level of LUT map. As the other limitations of the production and application of the LUT map can be considered as the necessity of computers for the research, software and color printing facilities.

5. Conclusions
In this research some of the new developments in digital mapping and GIS have been used to produce a new thematic map utilizing DEM and land use data. The study area was selected from southern Sri Lanka and DEM was generated manually and Land Use maps were produced by the digitally interpretation of Landsat MSS data. Ten these different data files were overlapped and finally produced a map which represents Land use and Topographic features simultaneously (LUT map). In the resulted map five land cover categories were included with eight topographic levels in different colors and tones. Apart from the creating LUT map; results generated a remarkable amount of information that can be deeply analyzed and cooperated with other data sources. As an example, the area calculation of the LUT map shows 836 of the are is below the 153 meter level and on the other hand it reports that it is only a 7.6% of the land goes under the Forest category. Further more the map explains 93% of the total paddy area spreads within land area between 0-152m. Also some limitations of the LUT map were observed as; a necessary of a scanner, computer, software, color printing facilities and limited number of classes that can be included in the color legend.

Acknowledgements:
Authors thanks are due to Dr. Koji Kajiwara (Dept. of Industrial Science, University of Tokyo) and MR. Hiroshi Okumura (Remote Sensing and Image Research Center, Chiba University, Japan) for their assistance and comments and to Mrs. M. Perera for the assistance given in data tabulation.

References

Burrough, P.A. (1986): Principles of Geographical Information Systems for Land Resources Assessment, Clarendon Press, Oxford.
Lillesand, T.M & Ralph, W.K.J. (1987): Remote Sensing and Image interpretation (Sec. Ed.) John Wiley & Sons. USA, Monmonier, M.S. (1982): Computer Assisted Cartography: Principles and Prospects, Prentic-Hall, Englewood Cliffs, New Jersey, USA.
Mather, P.M. (1991) : Computer application in Geography, John Wiley & Sons.
Perera, L.K. & Tateishi, R. (1991) : A Remote Sensing Oriented Guide Line to Environmental Management around Victoria Reservoir in Sri Lanka, Proc. Of the 12th Asian Conference on Remote Sensing (Vol. II), Singapore, Nov. 1991. p. J-3-1-6
Perera, L.K. & Tateishi, R. 91992): A thematic map based on Land use and Elevation data of Obitsu River Basin, Japan. Proc. of annu. Symposium of the Japan Society of photogrammetric and Remote Sensing, Tokyo, May 1992. pp. 83-88
Yapa L.S (1991): Is GIS appropriate technology? Int. Journal of Geographical Information Systems, Taylor & Francis, pp. 42-58.
Yeh. A.G.O. (1991) : The development and applications of geographical information systems for urban and regional planning in the developing countries. Int. Journal of Geographical Information Systems. Taylor & Francis, pp. 5-28.